Keywords: Regulation & Taxation; Social Policy;
During the writing of the 1984 Post-Election Briefing, Economic Management, Treasury economists got into a fierce argument over the purpose of taxation. Eventually a senior official, well known for his native wit, stepped into the heated fray and imposed the following opening of the relevant section of the PEB.
“The purpose of any tax regime is to raise revenue. The level of revenue will itself be dictated by the level of government expenditure and the size of the budget deficit that the government is prepared to accept.”
The briefing goes on to discuss how to design a tax regime, making the obvious point that not all taxes are equally effective. Alas, such common sense has not been so evident in the public debate on taxation in the 1990s. The focus has been on taxes, with nary a reference to the government spending it provides, even though – a theme of this book – New Zealanders have a high demand for public spending on retirement incomes, health, education, and a myriad of other activities. Perhaps because taxation is the downside of such spending, those with an agenda of reducing public spending want to emphasise only its disadvantages. This chapter cannot be so irresponsible.
The role of government spending
More than most modern societies, New Zealand society is the creation of the state, a perhaps inevitable consequence of its recent settlement. Public spending is one of the most powerful instruments which the state can use to create a civil society. Thus government spending has been an integral part of the New Zealand culture. Surveys show that New Zealanders support increased public spending, including that which does not directly benefit the respondents (as when the elderly urge more spending upon education), even if it means they pay higher taxation. In doing so, they affirm an involvement in a community: they are not just self-interested (and self-centred) individuals.
Undoubtedly there are many instances of advocacy of government spending (or other interventions) that are self-interested but presented as in the public interest. Teachers promoting more public spending on education are likely to improve their working conditions and remuneration if their advocacy is successful. Even so, one reason the advocates may be teaching – and perhaps having to accept inferior working conditions to those they could expect elsewhere – is that they believe education is socially important, and they are willing to take a financial sacrifice to pursue wider community goals. A common deception used to protect the self-interest paradigm is to argue that all behaviour is necessarily self-interested. Thus it is said the Good Samaritan really acted in his own interests when he helped the assaulted Jewish traveller. This verbal sleight on the meaning of ‘self-interest’ obscures the central point of the parable, namely that the Samaritan did not know who he helped (indeed, he belonged to a different community). Clearly we want to distinguish such behaviour from other kinds of behaviour where one is concerned only with oneself. It may satisfy the selfish to argue that everyone else is as selfish as they claim to be themselves, but common sense tells us this is nonsense. Each of us is a mixture of the self-interested and the public-spirited. The commercialisers choose to recognise only the mean-spirited component – but to do so is to reject the relevance of community to the human condition.
An additional complication is over-optimistic expectations. Advocates, committed to a policy for public-spirited or for self-interested reasons, will overestimate the benefits and underestimate the costs. Very often arguments will be seized upon to justify policies, even if they are wrong, irrelevant, or short-sighted. Thus the debate on government spending and taxation rarely has the cool rationality one might hope for.
A particular downside that is rarely mentioned by an advocate of public spending is that the additional taxation may result in costs on the economy. These are not just administrative costs, but include behavioural responses. Briefly, higher taxation may result in the taxed changing their behaviour in a way that is economically detrimental. Most commonly, they will reduce the activity which is taxed, to some degree. If that activity is labour, the economy may have less output; if it is consumption, total spending on the commodity may be below that which is valued by consumers.
There is an understandable tendency to use a simplified tax-free economy as a reference point to examine the effects of taxation on spending. However, the reference point is but a benchmark, rather than an economy with ideal properties. Treating the tax-free economy as some paragon is about as sensible as treating a temperature of absolute zero as special. Indeed, the point of advocating government spending is that the output of a tax-free (and hence public-spending-free) economy is criticised as not the best available. Those who are pressing for (more) public spending on education, health, the arts, or whatever it may be, are arguing that the output of the tax-free economy is inferior to one in which this public spending occurs. If the public judges them correct, it is illogical to use the output of the tax-free economy as a measure to evaluate the effects of the taxation and public spending. GDP may not be the best measure of economic welfare. The public appears to value much government spending more than it is valued in GDP.
Unfortunately, this gives little guidance as to how much public spending there should be, or what it should be spent on. That, ultimately, is a political question: some politicians and parties may argue for reductions in spending, others may argue for more. In practice, there is a need to judge each spending proposal in terms of its social benefits (realistically, not optimistically) and any offsetting costs (including the behavioural impact of the additional taxation). Typically, the government’s chief economic adviser, the Treasury, faces enormous pressures for additional spending and for reducing taxation: among government advisers it faces such pressures alone. A cabinet might be thought of as nineteen ministers (backed by their caucuses) who want to increase spending, and a lone Treasurer who is expected to fund the increases (usually without raising taxation). Hence the Treasury’s support for fiscal austerity, although it does not necessarily lead to the ideology of commercialisation adopted in the mid-1990s.
From the introduction in 1988 of the Public Finance Act, and especially in the 1990s under the National Government, there has been a steady squeeze on public spending. Much of this book is about the consequences of that squeeze on the state and on society. However, in the mid 1990s the political-economic debate tended to focus on levels of taxation without reference to spending. The debate may well have happened anyway, but much was institutionalised via the Department of Inland Revenue (IRD). In order to restrain the rampant Treasury minister Ruth Richardson, the National Government attempted to build up the policy advice capacity of the IRD under its minister Wyatt Creech, who sometimes bitterly clashed with her. However the IRD’s hurriedly created policy advice arm seems to have got out of control, in the sense that its resources far exceeded the quality of its work.
The taxation debate (Diewert & Lawrence)
But poor-quality research on the effects of taxation was not confined to the IRD. A Business-Roundtable commissioned report, if taken seriously, suggested that the taxation reforms of the 1980s had been disastrous. The study by two overseas consultants concluded that the last dollar (i.e. the marginal dollar) spent by government ‘means forgoing $1.18 of benefit that would otherwise accrue to taxpayers. If the dollar of government spending were worth a dollar to the taxpayer, the gain from reducing government spending would be 18 cents.’ While the report’s introduction described the loss as ‘high’, the 18 percent figure (if true) means that if government spending at the margin was valued by the public at about a fifth more than it was valued in GDP terms, the taxation was worthwhile. This interpretation was completely ignored, even though it may more accurately reflect New Zealanders’ desires than the views advanced by the Business Roundtable.
The tax regime mentioned above was that of 1990, but the consultants also calculated the figures for earlier years. Before the reforms of the 1980s, the deadweight loss was 5 to 7 percent. Thus it appears that the effect of the reforms was to more than double the costs to the economy of income tax.  Apparently the tax regime under Rob Muldoon was more than twice as efficient as that introduced by Roger Douglas, even though the regime in Muldoon’s time was widely criticised – even by the government-appointed 1982 McCaw Committee – for its arbitrary exemptions and its complexity. Most independent expert commentators commend the Douglas reforms for creating a simple broad-based income-tax regime which rationalised expenditure taxation. (Many criticise the Douglas reforms for favouring the rich against the poor. That is a different issue from the overall design of the tax system. Most of the features of the Douglas system could have been retained with a higher marginal tax rate on top incomes.)
But there is a gross flaw in the consultants’ report, invalidating these conclusions. To simplify a little, the deadweight losses from a tax on labour earnings arise from the behavioural response of choosing to work less because the after-tax return is lower than before-tax return. Workers stop working because the taxes on their extra earnings do not make it worthwhile to labour. Instead of earning, and adding to the production of the nation, they choose more leisure. To make their model work, the consultants had to decide how much additional leisure there was in the economy as a result of the tax changes. They seem to have attributed all the increase in unemployment in the 1980s to higher tax rates. The unemployment rate more than doubled between 1984, when Muldoon left office, and 1990. In effect, the consultants claimed that all the additional unemployment is voluntary. They assumed that if only taxes were lowered, the unemployed would stop their ‘leisure’ activities and take up work. However, most of New Zealand’s unemployed do not have that choice – they simply cannot find a job. Rather than being discouraged by high tax rates, they are involuntarily unemployed. Thus it is unlikely that the Douglas tax regime was twice as inefficient as Muldoon’s which preceded it. The later regime – distributional objectives aside – is almost certainly more efficient.
The taxation debate (Scully)
The IRD-commissioned studies were even stranger. It is said that the department spent over $2m on its various commissioned studies. There have not been the resources outside the IRD to evaluate all the studies. We look at only one (probably costing around $100,000), but if the most prominently quoted one is so flawed, despite the quality control the department claimed it had, we must have similar reservations about all the others – until they have been properly and independently evaluated.
US economist Gerald Scully estimated the (‘optimum’) level of taxation which would maximise the GDP growth rate.  It would be tedious to go through all the weaknesses of the study, although they alert the experienced econometrician that there is a problem. For the layperson, it is important to note that correlation does not prove causality (although a lack of correlation means there can be no causal relationship). Scully estimated that an ATR (average tax rate; the ratio of tax revenue to GDP) of about 20 percent would give a maximum growth of volume GDP. The figure has been seized upon by politicians of a right-wing political persuasion, especially those in Act. In 1999 the actual ATR was about 31 percent for central government taxation, so the advocates are covertly arguing for a major cut in government spending.
Note, incidentally, that Scully’s work says absolutely nothing about government spending or the government deficit. It appears to say that a 20 percent ATR would maximise economic growth, whatever the level of spending or deficit. Taking the research at its face value, one could advocate an ATR of 20 percent of GDP and a government spending rate of 31 percent and apparently obtain the maximum long-run optimal growth rate of volume GDP, despite the size of the government deficit that would be involved. Note also that the measure takes no account of the design of the tax regime, treating the clumsy Muldoon one as equally effective as the elegant Douglas one. Indeed, one of the oddities of the IRD research programme is that it did not address the area where the department had the greatest expertise and interest: the design and administration of an effective tax system. Instead, it examined an issue outside its remit: the relationship between taxation and macroeconomic performance.
Because the main fallacy of the Scully work involves some mathematical and statistical understandings, I have relegated the exposition to an appendix (not included here but in the book). The non-mathematical/non-statistical reader can see the work is silly, because Scully’s method produces absurd results. If there is an ATR which maximises economic growth, we might ask what ATR maximises other social phenomena. So I replicated Scully’s work, but I also applied the method to the growth of deaths and of reported criminal offences. The results are summarised in the following table. In both cases, the maximum growth rate is given by an ATR of about 20 percent of GDP. Apparently high economic growth is associated with high mortality and high criminality. A Hobbesian outcome!
Optimal tax rate for various variables using Scully’s econometric method
|Variable to be
|ATR for maximum growth
(Proportion of GDP)
The book table also includes the 95 percent confidence interval for each estimate. All cover .2
As the appendix shows, the results are the consequences of ‘spurious correlation’, induced by the way the data has been managed. A heuristic explanation of the results is that the actual econometric equation involves a transformation which results in the ATR being correlated on a (curved) transformation of itself. Providing the other variable in the equation is not correlated with the ATR (and thus there is no causal relationship at all between the ATR and the variable which is being allegedly maximised), the econometric equation will give an apparent maximum related to the average ATR. The 20 percent ATR conclusion only applies when the hypothesis being assumed is not true. The absurdity is well illustrated by using as the independent variable the number one (i.e. 1 for every observation). As the last row in the table shows, the econometrics dutifully advises that an ATR of about 20 percent maximises the number 1. Which is nonsensical: a change in the tax rate does not change the value of a number.
Economic research and ideological politics
Science proceeds by such mistakes, to be corrected by later scientists. Normally one would expect this to happen here (although taxpayers might be disgruntled over the amount they paid to fund it). However, social sciences often function differently, in that work will be seized upon for ideological purposes, even if it is not scientifically robust. It is therefore worth exploring the wider context in which the Scully work was done. When I published a simplified account of these findings, numerous economists said they agreed with me, or had come to a similar conclusion that the result was based on spurious correlation. Although Scully and the IRD chief tax policy adviser at the time of the work, Patrick Caragata, denied the finding, they did not try to replicate my analysis, and continued to promote the Scully result.
The inadequacy of much economics research derives from at least three features. First, many economists have a very shallow methodology, deriving from the tight prior approach which sees the task of empirical research as being to find support for the theories which are tightly held, rather than testing them to see whether they are consistent with the real world.Thus Scully seems to have had a theory that tax rates in New Zealand were too high. Since his econometrics supported that theory, it was uncritically accepted. In contrast, a scientist considers whether there are alternate theories (such as spurious correlation) which fit the evidence better.
A second feature is the dependence on ‘fifos’, fly-in-fly-out overseas consultants who spend a short time in the country to carry out work for government agencies (or the Business Roundtable). Why are the agencies so dependent upon fifos, who are not very knowledgeable about the New Zealand context of the problem they are looking at, and often not very competent, as shown by the quality of their work? Why not use competent New Zealand economists (as shown by their ability to provide penetrating critiques of the fifos’ work if the opportunity arises), especially as this would build up the general competence of the New Zealand economics profession? There are probably two reasons for the limited use of New Zealand economists. First, the fifo can be controlled by the agency. Fifos may be independent, but the one with the right ideological bias can always be selected, and the consultant’s ignorance of New Zealand can be relied on to ensure that the agency’s assumptions are not challenged. Some would argue a second reason is the need for ideological control by the agency, but the issue is probably more subtle than that. Many economists in government agencies seem unwilling to go to New Zealand economists more competent than they, in case their own ignorance might be exposed. Moreover, a consultancy can increase the competence of the New Zealand economist, which one day might be turned against the agency. Better to go to the fifo for a one-night stand.
The third feature is the low quality of the reviewing of the research work. It is a standing joke that not to be invited to some presentations of commissioned research is evidence of one’s competence, for if a skilled person was asked to review the work, both the consultant and the agency which commissioned it would be exposed. Since the study meets the ideological requirements, if not the scientific ones, the only outcome of a rigorous review would be to undermine the research and raise questions about the wasting of funds. In defence, Caragata says that the IRD work was refereed by some economists the IRD chose, and it was published in an overseas journal. Others will conclude either that such refereeing was superficial, or else that the referees were not very knowledgeable about the particular research topic.
In an ideal world, scientists would offset this ideological predominance. In the natural sciences and medicine, they continue to do a reasonable job in New Zealand. But the closer the science relates to core political issues, the harder the task. Given the undermining of the independent scientific process in New Zealand, it commonly happens that politicians will abuse research for ideological reasons.
The future of the taxation debate
There will always be advocates of lower taxes, and the 20 percent target (central government’s tax take as a proportion of GDP) may be locked into the New Zealand political debate for some time. One would like to say that the advocates will honestly and as frequently state that they are advocating equally substantial reductions in government spending, which in practice would involve major privatisation in health and education, and reductions in social security benefits, together with a substantial distributional shift in favour of the rich. One might even hope that the advocates will eschew poor-quality economic research which seems to support their political objectives. More likely, they will focus on lower taxes, ignoring any spending and distributional consequences, using inadequate research when it suits them.
The other side of the debate can be equally cavalier. Interest groups will advocate government spending increases which contribute to their causes, usually without mentioning the need for higher taxation, and will quote equally faulty research when it suits them. It is more difficult to predict whether there will be a strong lobby for more government spending across the board and for the consequent tax increases. Both Labour and the Alliance are tempted towards this approach (albeit typically without mentioning the tax consequences), and there are even signs of support on the left of the National Party (i.e. the centre right), thus reflecting the stance of the majority of the public.
Even so, political parties seem to be reluctant to advocate increased public spending very strongly, perhaps for two reasons. First, while those who want lower taxes may be in a minority, theirs is the more vigorous political rhetoric. Second, there is the problem of finding the tax revenues to complement the spending. There may be some political mileage in a higher top tax rate on personal incomes. As Chapter 4 reported, both the Irish and Australian economies, which have had a better economic performance than New Zealand, also have higher top tax rates. However, raising the top tax rate generates only a little revenue in comparison to the spending demands. Among other extensions which might improve the effectiveness of the tax regime, and raise a little revenue, are (1) a real capital gains tax to prevent distortion of investment and savings decisions; (2) a GST on various currently uncovered foreign consumption transactions (including tourism and international electronic commerce); and (3) a financial transactions tax. However, the big gains for improving the tax regime were made in the 1980s under Labour.
Which leaves the advocates of public spending largely arguing that the spending should increase in line with increases in GDP.
(The Appendix is Omitted)
1 The crucial notion here is `regime’. Individual taxes may have other purposes, such as to align private costs with social costs. (Hence the excise duties on licit drugs.) Treasury (1984:210)
2. Easton Listener 1 March 1996, p.630 May 1996, p.75; RCSP (1988); NBR, 26 Feb 1999.
3. Part of the problem is that the commercialisers, besotted with US economic theory, do not recognize that in a federation as large as the US, taxation is less likely to benefit the community of the taxpayer. (Hence the strong pork barrel politics in the US congress.) This is another example of the fallacy which persists in commercialist thinking – the US is a model economy and all others should be forced to conform to it.
4. If the commodity is an economic bad – such as pollution – the reduction in consumption would be a good thing. However, the totality of economic bads is insufficient for taxes on them to fund all the economic goods that is desired. In any case, given an objective of the complete elimination of most economic bads, the taxation revenue on the (almost) eliminated is going to be (almost) zero.
5. I do not deal in detail the issue of the size of the government deficit, which temporarily allows spending to be funded from non-revenue sources. Briefly, a deficit adds to public debt, which has to be serviced from future revenue, so that a deficit shifts forward the required taxation rather than permanently avoids it. To some extent a deficit may contribute to economic growth, and pay for itself from the higher government revenue, as many Keynesians would argue. But this occurs only to a limited extent. Once that limit is exceeded, the situation in the second sentence applies, either immediately or in the future.
6. CONZ (85-98)
7. Richardson (1995); Creech (1996)
8. Diewert & Lawrence (1994)
9. The consumption taxes show a similar difference.
10. Sieper (1997) makes some similar points.
11. The Treasury has distanced itself from this IRD research. (Sieper 1997) When it is not down its ideological path, Treasury remains the most competent of all government economic agencies. The IRD work was so flawed, and had such horrendous policy implications – such as the ignoring the government deficit and not designing an effective tax regime – that the Treasury needed to reject the IRD findings.
12. Scully (1996, 1997)
13. This includes the income taxes on benefits and GST on government spending, which increase the ratio in comparison to most other OECD economies.
14. My estimates differ from Scully’s at the third significant figure, presumably as a result of rounding errors. I have used the Scully ATR, even though it does not include all the taxation revenue received by the government. John Lepper has pointed out that unemployment levies in the 1930s did not go to the consolidated fund, and it seems likely that other such revenues were also overlooked.
15. Another demonstration of the inadequacy of the econometric equation, is that if it is used to predict the growth rate, the error of prediction is greater than if the average had been used.
16. Listener, 12 July 1997.
17. Caragata (1997, 1998) The IRD asked me to provide a report but were unwilling to pay a fee for it. They did not even discuss a contingent fee, that is I would only get paid if the work was of acceptable quality. A number of agencies seem to have a belief that if they pay good money for a low quality work, they should get high quality critiques of it for free. Caragata misrepresents this interchange in his 1997 report.
18. CONZ (91-94); Reder (1982)
19. In 1987 I attended a lecture in Australia, where an Australian economist who had worked for the New Zealand Treasury on labour relations (so he said) told the audience that it was impossible to get a newspaper in New Zealand before 10.00 in the morning because of union restrictions.
20. 45 percent and 47 percent respectively in contrast to New Zealand’s 33 percent (although Labour has promised to raise it to 39 percent for incomes above $60,000 p.a.).